Sharp-tailed grouse Tympanuchus phasianellus were effectively extirpated from western Montana during the last century as a result of settlement by Euro-Americans. Recent interest in reestablishing the species west of the Continental Divide has identified information gaps related to the potential success of a restoration effort. Elsewhere, sharp-tailed grouse are widespread and exhibit plasticity in habitat use, suggesting a high potential for successful reintroduction. Using life history information from the published literature, we conducted a population viability analysis to assess the potential viability of a reintroduced population of sharp-tailed grouse in western Montana and to evaluate what management scenarios, with regard to both translocation protocols and habitat management, would be necessary to produce a viable population. Results of the population viability analysis indicated that a population parameterized with mean reported demographic rates and related environmental variation would not be viable and suggest a potential downward bias in demographic estimates in the published literature. Based on our simulation results, improvements in both fecundity and annual survival resulting from improvements in nesting and winter habitat would be necessary to produce a viable population of sharp-tailed grouse in western Montana. The minimum amount of habitat required to support a viable population of 280 individuals was 1,867–5,600 ha, assuming habitat is sufficient to support an average density of 5–15 grouse per km2. We provide a review of demographic and reintroduction information for sharp-tailed grouse and recommendations regarding reintroduction approaches based on our population viability analysis results that should increase the relative success of restoration efforts in western Montana and elsewhere. We recommend that nesting and winter habitat improvements be the focus of pre- and postrelease management and that post-translocation population studies be conducted to monitor reintroduced populations and provide site-specific demographic information to update population viability analyses.
An important upland game bird throughout much of its range, sharp-tailed grouse Tympanuchus phasianellus historically occurred throughout Montana, including in grassland-dominated mountain valleys west of the Continental Divide (Silloway 1901). Prairie-grouse (genus Tympanuchus) populations have declined throughout their range over the last century due to degradation, fragmentation, and loss of habitats (Aldridge et al. 2004), and western populations of sharp-tailed grouse in Montana declined rapidly with Euro-American settlement of the region. Although sharp-tailed grouse populations are currently widespread and stable in eastern Montana, populations west of the Continental Divide are likely extirpated (Fitzpatrick 2003; Young and Wood 2012).
Sharp-tailed grouse occupy diverse grassland, steppe, and mixed shrub ecosystems throughout central and northern North America and, as a result, are thought to tolerate a greater variation in plant community types and composition than other species of prairie-grouse (Johnsgard 2002). Plasticity in habitat use suggests that sharp-tailed grouse may have a high potential for successful translocations and reintroductions. Nevertheless, with large home ranges and differing requirements for nesting and winter habitat, sharp-tailed grouse require large and complex areas containing a variety of vegetation types (Sisson 1976; Marks and Marks 1988; Giesen and Connelly 1993; Connelly et al. 1998; Goddard and Dawson 2009b). As habitat quality and quantity interact to affect the demographic performance of wildlife populations (Hobbs and Hanley 1990), the availability and appropriate juxtaposition of key seasonal habitat components should be considered when selecting potential restoration sites. Reproductive potential is typically high among prairie-grouse due to high rates of nesting, large clutch sizes, and high hatching rates (Bergerud and Gratson 1988). Population sensitivity analyses of sharp-tailed grouse and lesser prairie-chickens Tympanuchus pallidicinctus concluded that reproductive success, including both nest and brood survival, had the largest impact on population dynamics, with female survival playing a lesser role (Hagen et al. 2009; Gillette 2014). However, the timing and severity of weather events can also significantly influence overwinter survival and thus recruitment of birds to spring breeding populations (Ulliman 1995). Based on these results, wildlife managers can improve population performance by focusing first on nesting and brood-rearing habitats, followed by winter habitats.
Translocations have frequently been used to restore or supplement declining populations of prairie-grouse (Snyder et al. 1999). However, translocations, and reintroductions as a whole, have been characterized by a lack of monitoring or standard metrics for success, thereby reducing their utility to inform future efforts (Ewen and Armstrong 2007). Prairie-grouse, in particular, have proven difficult to restore to historic habitats, and a lack of documentation of previous efforts has limited the understanding of factors related to success (Toepfer et al. 1990; Snyder et al. 1999).
Population viability analyses (PVAs) are a common tool used to aid decisions when managing wild populations, particularly with species of conservation concern (Beissinger and Westphal 1998). Standard uses of a PVA include estimating 1) extinction probabilities for one or more populations over a set time period, 2) a minimum viable population (MVP), and 3) the minimum dynamic area (MDA; Reed et al. 2003). The MVP is the number of individuals required for a population to have a given probability of persistence over a specified period, whereas the MDA represents the smallest area of ideal habitat required to support the MVP. Although these metrics can provide baseline information, an optimal use of a PVA is the comparison of different management options to evaluate the relative differences in extinction probability compared to a baseline scenario (Beissinger and Westphal 1998). In this regard, PVAs can augment knowledge from previous translocations by comparing different translocation and habitat improvement scenarios to inform future efforts.
There is currently interest from Montana Fish, Wildlife and Parks and conservation organizations in restoring sharp-tailed grouse populations to western Montana. To evaluate the potential for success and inform restoration efforts, we conducted a PVA for sharp-tailed grouse by using demographic and life history information in the published literature. Our objectives were to 1) determine whether a self-sustaining population of sharp-tailed grouse is possible in western Montana; 2) identify the MVP for sharp-tailed grouse by using the best available demographic information and stochastic population modeling; 3) identify the MDA required to support a sustainable population of sharp-tailed grouse; and 4) compare the relative effects of management scenarios, with regard to both translocation protocols and habitat management, on the probability of restoration success.
We used program VORTEX 10 (Conservation Breeding Specialist Group, Apple Valley, MN) to conduct PVAs and evaluate hypothetical management scenarios for the reintroduction of sharp-tailed grouse into western Montana. The program VORTEX is an individual-based Monte Carlo simulation package that can model the effects of deterministic population parameters as well as demographic stochasticity, environmental variation (including catastrophes), genetic stochastic events, and intrinsic population regulation (Lacy 1993), and it has been used to evaluate population viability for many species (Lacy 2000). Population processes are modeled as discrete, sequential events by randomly drawing individual parameter estimates from user-specified distributions to simulate births and deaths of individuals that are tracked throughout the length of each simulation (Lacy 1993).
We modeled a single population without immigration and projected 1,000 population trajectories over 50 y for multiple management scenarios. A time period of 50 y is relevant to management, has been used in other PVAs of isolated sharp-tailed grouse populations (Temple 1992), and reduces the effects of error propagation and the assumption that the system is indefinitely static (Pe'er et al. 2013). We used a prebreeding population model where the sequence of events was 1) environmental variation, 2) breeding, 3) mortality, 4) ageing, 5) supplementing, and 6) census. We set initial population size at 75 individuals (25 males, 50 females) based on a trade-off between feasibility and cost of translocating prairie-grouse (Snyder et al. 1999). The initial sex ratio of the released cohort was based on maximizing the number of reproductive females and the assumption that an initial cohort of 25 males would be released the prior fall to establish lek sites and minimize movements away from restoration sites (Rodgers 1992). Initial modeling suggested that adjustments of initial abundance, sex ratio, or both did not significantly affect simulated results (L.B. McNew, unpublished data). We defined a viable population as one with a 95% probability of persistence for 50 y (Temple 1992). We considered the probability of extinction as the proportion of simulated populations within a given scenario with only one sex remaining after 50 y. We examined the intrinsic rate of population change (r) for each scenario to evaluate population trends in addition to the probability of extinction.
Program VORTEX requires a description of the initial population including the species' reproductive strategy, rates of inbreeding depression, carrying capacity (K), and parameter rate distributions for both fecundity and annual mortality, which can vary by age or sex. We estimated parameter distributions for our population model using the best available information from published literature on prairie-grouse demographic rates (Tables 1 and 2). Demographic stochasticity is the random variation in observed vital rates due to sampling and is incorporated into models within VORTEX through the random determination of whether a specific individual lives and breeds. In addition, we specified estimates of environmental variation around vital rates, which reflect fluctuations in the probabilities of births or deaths due to random changes in the environment.
We modeled a polygynous mating system with 20% mate monopolization (Robel et al. 1970; Gratson et al. 1991). We set the maximum lifespan of birds to be 7 y and assumed that both females and males can breed at 1 y of age and that prairie-grouse can reproduce until death (Gratson et al. 1991; Connelly et al. 1998). No density dependence in either reproductive effort or reproductive success was considered (Bergerud and Gratson 1988; Wisdom and Mills 1997; Roersma 2001). We assumed that all females reproduced under normal conditions, but we reduced the probability of nest initiation by 52% in the first year post-translocation (Coates et al. 2006). As demographic rates can vary considerably due to normal annual variability in weather, habitat conditions, predation, and disease, we conservatively parameterized environmental variation in fecundity to be 10% based on published rates (McDonald 1998; Moynahan et al. 2006; Hagen et al. 2009; McNew et al. 2012). Sharp-tailed grouse only produce one brood per year, and we assumed a maximum clutch size of 17 and a 50:50 sex ratio at hatch (Connelly et al. 1998). We calculated fecundity (F), or the number of fledglings produced per female at 14 d posthatch, as a function of parameters from the published literature by using the following equation:
where NEST represents the percentage of females that attempt a nest; CS1 and CS2 represent the clutch sizes of first and renest attempts, respectively; NSUCC1 and NSUCC2 represent nest success of first and renest attempts, respectively; RENEST is the probability of renesting; CPE is the number of chicks produced per egg laid; BSUCC is brood success; and FPC is fledging success at 14 d posthatch (McNew et al. 2012). Nest success is defined as the probability that a nest produced one or more chicks, and brood success is defined as the proportion of broods that successfully produced one or more chicks at 14 d posthatch. We used bootstrapping procedures to calculate estimates of error for fecundity by randomly drawing from the underlying distributions of input parameters by using Program R (R Development Core Team 2014). We modeled fecundity as ∼Poisson (3.5) in the baseline model as calculated from initial vital rates and related measures of error derived from reported rates in 15 published studies (Table 1).
We set annual survival to 0.50 based on the mean value from a review of 10 published studies (Table 2). However, annual survival of translocated female sharp-tailed grouse was reduced by 50% in the first year post-translocation (Mathews et al. 2016). Based on these estimates, we set annual survival of translocated birds at 0.25 for the first year. We modeled density dependence in survival by setting a threshold K such that survival is truncated when K is exceeded (see below). Information on juvenile survival is lacking for sharp-tailed grouse; therefore, we used published rates from the literature and set average juvenile survival, from 14 d posthatch to the following spring, to 0.40 based on three studies of greater sage-grouse Centrocercus urophasianus, greater prairie-chicken Tympanuchus cupido, and lesser prairie-chicken (Beck et al. 2006; Pitman et al. 2006; McNew et al. 2012). Variation in annual survival rates of prairie-grouse can be significant and has ranged from less than 5% to more than 50% (Moynahan et al. 2006; McNew et al. 2012; Davis et al. 2014). We set annual environmental variation in survival rates to 15%, which we assumed to represent typical annual variability in survival rates (McNew et al. 2012).
In addition to the standard annual variation in demographic rates due to environmental stochasticity, we considered two types of potential catastrophes with different probabilities of occurrence. A catastrophe is a type of environmental variation that is not necessarily rare but is distinct from environmental stochasticity due to its extreme effects on demography that result in large population declines (Beissinger and Westphal 1998). First, we modeled an extreme winter of annual snowfall exceeding the annual average by 1 standard deviation with a 0.02 probability (one extreme winter in west-central Montana during the past 60 y based on National Oceanic and Atmospheric Administration data). We estimated survival during the extreme winter to be 0.34 compared to a normal year (0.5) based on differences between survival rates in mild and severe winters for sharp-tailed grouse (Ulliman 1995). The second type of catastrophe that we considered was a cold wet spring and summer with mean temperatures and total precipitation that exceeded the standard deviation, which we modeled to have a 0.06 probability of occurrence based on National Oceanic and Atmospheric Administration data reported in the past 37 y; we parameterized this catastrophe to reduce fecundity by 34% (Erikstad and Andersen 1983; Smyth and Boag 1984; Bousquet and Rotella 1998).
Genetic health, including heterogeneity and allelic diversity, is a particularly important consideration for small populations. Loss of genetic diversity can facilitate inbreeding depression, potentially resulting in lower fecundity and resistance to disease, and a population that is less capable of adapting to changing environments. Moreover, loss of genetic diversity can hasten population extirpation through a positive feedback loop known as the “extinction vortex” (Gilpin 1986; Allendorf and Ryman 2002). The loss of genetic health in prairie-grouse populations has been linked to a decline in egg hatchability (Bouzat et al. 1998), but the rates of drift and other stochastic genetic events are unknown for prairie-grouse. Therefore, we used default values and modeled stochastic genetic effects on fecundity and first-year survival by using the average value of 6.29 lethal equivalents reported for 11 species of vertebrates in the literature (O'Grady et al. 2006).
We incorporated and evaluated the effects of different levels of habitat quality or quantity by testing scenarios at varying levels of K as a proxy for the amount of available habitat. The VORTEX program does not allow users to distinguish between quality and quantity of habitat. K is implemented by truncating population growth when K is exceeded through additional mortality, such that population size equals K after truncation. Although little to no information is available on sharp-tailed grouse population densities in the winter, breeding densities have typically varied from 0.6 to 5.5 birds per km2 (Connelly et al. 1998), although Hamerstrom (1939) reported a density of 25 birds per km2. Based on an average density of 15 birds per km2 in ideal conditions, we ran each scenario at K values of 500, 1,000, 2,000, and 4,000 individuals to model the effects of a range of habitat sizes on population trajectories by using population size as a proxy for habitat availability.
We adjusted vital rates of our baseline model to evaluate 10 hypothetical management scenarios based either on translocation methods or habitat management that are likely to be used by Montana Fish, Wildlife and Parks and other conservation groups as guidelines for future sharp-tailed grouse reintroductions in western Montana (Table 3). Unless otherwise stated, we modeled all management scenarios at the four levels of K described above. Three scenarios modeled different translocation techniques. First, we modeled translocating just yearlings (scenario A) based on reported 59% higher survival (0.67 vs. 0.42) of yearling vs. adult translocated sharp-tailed grouse (Mathews et al. 2016). Second, we modeled the effects of supplementation on population persistence by supplementing 10 grouse annually for the first 5 y at K values of 500 and 1,000 for the baseline (scenario Ba) and juvenile (scenario Bb) translocation scenarios. Third, we modeled a genetic rescue by supplementing 10 grouse every decade for the baseline (scenario Ca) and juvenile (scenario Cb) translocation scenarios (Mills and Allendorf 1996).
Six additional scenarios modeled the effects of different habitat management actions. Based on previous sensitivity analyses, we focused on habitat management actions that would improve either reproductive success or adult survival (Hagen et al. 2009; Gillette 2014). First, we modeled the effects of predator removal for the first 2 y of translocation effort (scenario D). For greater sage-grouse, raven Corvus corax removal increased nest survival by 73% (Coates and Delehanty 2004); the adjusted nest survival rate corresponded to an estimated increase in mean fecundity to 4.7 offspring per female. We did not include mammalian predator impacts on nest survival because previous research found no effect of mesopredator trapping on nesting success of ground-nesting birds (Wiens 2007). A second habitat management scenario involved two alternative methods to improve nesting habitat either through the removal of conifers (scenario E) or improved grazing practices (scenario F), both of which improved reproductive success. Nest survival increased by 5.2% for a population of greater sage-grouse with the removal of conifers (Severson 2016) and by 37% without heavy grazing in a population of black grouse Tetrao tetrix (Baines 1996). We modeled both effects on fecundity separately. We also modeled the effects of winter habitat improvement as a 15% increase in shrub cover (scenario G), which increased overwinter survival by 19% in a population of greater sage-grouse (Moynahan et al. 2006). Finally, we modeled improvements in both nesting and winter habitat by combining the increased overwinter survival due to greater shrub cover and the increased nest survival either due to conifer removal (scenario H) or due to improved grazing practices (scenario I).
After examining both translocation and habitat management scenarios, we then manipulated the top scenario to determine the MVP by decreasing the K to the minimum that allowed for a population with a 95% probability of persistence for 50 y. We also evaluated a combination of the best translocation and habitat management scenarios by examining supplementation of 10 females every 10 y with both improved overwinter shrub cover and improved grazing practices (scenario C + scenario I = scenario J).
Finally, we conducted perturbation analyses to evaluate the impact of using parameters from species other than sharp-tailed grouse on our results. We varied estimates of juvenile survival in the baseline model and estimates of adult survival and nest survival (and therefore fecundity) in the top model to determine the range of parameters over which our evaluation of a viable population remained consistent. We considered our results robust if they were not altered under a range of possible parameter estimates that exceeded the coefficient of variation for that parameter (Tables 1 and 2).
Viability and minimum viable population
Simulation results indicated that a viable population was only achieved under scenario I with increases in fecundity and overwinter survival realized through improvements to breeding season and winter habitats (Table 4; Figures S1–S3, Supplemental Material). Following these results, we reduced K for scenario I to test smaller population sizes of 200 and 280 individuals, both previously suggested to be sustainable (Toepfer et al. 1990; Temple 1992). This yielded an MVP of 280 sharp-tailed grouse in a landscape managed to improve sharp-tailed grouse vital rates over average rates reported in the literature. Under scenario I, the average rate of population increase over a 50-y run was 23% (r = 0.23). The population increased rapidly toward K (280) during the first 18 y after introduction. However, average population size was predicted to decline slowly and the predicted size of the average simulated population at year 50 was approximately 228 birds. The population size of 200 recommended by Toepfer et al. (1990) and Hoffman et al. (2015) had a 93% probability of persistence for 50 y under scenario I and was not considered a viable population by our a priori criteria. Increasing K generally improved population persistence and reduced the amount of genetic diversity lost in each scenario. However, increasing K did not lead to a viable population in any scenario except in scenario I. Larger K values under scenario I also produced a population with a 95% probability of persistence for 50 y and maintained a larger proportion of the introduced cohort's genetic variability. The MVP of 280 individuals under scenario I lost approximately 20% of its initial genetic diversity over the 50-y period, whereas populations in areas with higher K under the same scenario lost approximately 10–15% of the initial genetic diversity (Table 4). The genetic effects on small populations negatively affected the viability of the population at K = 200 under scenario I, whereas a population of 280 individuals was sufficiently large enough to mitigate the negative effects of inbreeding and genetic drift.
Minimum dynamic area
The MDA for a viable population of sharp-tailed grouse is a function of the K and was estimated to be 1,867 ha for a maximum population size of 280 individuals, assuming the birds can be sustained at a density of 15 birds per km2 (Hamerstrom 1939). If larger populations of 2,000 or 4,000 sharp-tailed grouse are desired to prevent the deteriorating effects of a loss of genetic diversity on small populations, we estimated MDAs for these larger populations to be 13,333 ha and 26,667 ha, respectively. An average density of 5 birds per km2 may be more realistic (Table S1, Supplemental Material); under this lower average density, a population of 280 birds would require 5,600 ha of suitable habitat.
None of the original scenarios of translocation and supplementation (baseline, A, B, and C) resulted in a viable population. Although early population supplementation maintained short-term genetic diversity and increased the population size for the first 10 y, overall genetic diversity decreased 20–30% and the populations declined on average (r < 0; Table 4; Figure S1, Supplemental Material). Similarly, translocation of yearlings increased the size of the population initially relative to the baseline scenario, but the population decreased by 10% on average and approximately 20–30% of genetic diversity was lost (Table 4; Figure S1, Supplemental Material). Supplementation of 10 females every 10 y was not sufficient to produce a viable population, but under the MVP scenario (scenario I + scenario C = scenario J), supplementation increased long-term genetic diversity and increased the population's probability of persistence (Figures 1 and 2; Table 4). Further scenarios combining habitat improvement and supplementation scenarios (e.g., nesting and winter habitat improvements + supplementation of 10 females each year for the first 5 y) could be explored with an aim to maintain genetic diversity and alleviate concerns related to genetic effects on small populations.
Simulation results indicated that the most limiting factors to population persistence were depressed fecundity and winter survival that are likely mediated by nesting and winter habitat conditions (Table 4). Improvements in fecundity associated solely with predator removals in the first 2 y post-translocation were not sufficient in producing a viable population of sharp-tailed grouse during our simulations. Under this scenario, population size and persistence were only slightly higher than those of the baseline scenario and the population decreased by approximately 9% on average. All habitat improvement scenarios except conifer removal produced a population that increased on average (r > 0), but not all populations persisted for 50 y. Of the scenarios with a single habitat improvement action (E, F, and G), the improved shrub cover scenario (G) had the highest probability of population persistence (82–83%) and maintained the highest genetic diversity (12–18% loss), whereas the improved grazing practices scenario and the conifer removal scenario had a probability of population persistence of 38–48 and 12–17%, respectively, and a genetic diversity loss of 14–25 and 19–28%, respectively. Expected demographic increases resulting from single habitat management actions and a combination of conifer removal and increased shrub cover (scenario H) increased population persistence relative to other scenarios, but did not produce a viable population. However, a combination of improved grazing practices and winter shrub cover improvements (scenario I) resulted in expected increases in fecundity and annual survival that resulted in a viable population of sharp-tailed grouse (Figure 3).
Perturbation analyses suggested that the results of our baseline model were robust to changes in juvenile survival; our conclusions regarding population persistence did not change under a range of 0–80% juvenile survival. Our perturbation analysis of the top scenario (scenario I), in which we varied estimates of both fecundity and adult survival, suggested that our assessments of population viability were more sensitive to small changes in adult survival than changes in nest survival. However, our conclusions regarding population persistence did not change under a range of parameter estimates for both adult survival and fecundity that exceeded the coefficient of variation among published studies (Figure 4).
Based on our simulation results using vital rates reported in the literature, significant improvements to both nesting and winter habitat are necessary to produce a viable population of sharp-tailed grouse in western Montana. If habitat improvements lead to increased fecundity and overwinter survival, a self-sustaining population may be possible with a minimum viable population of 280 individuals and a minimum dynamic area of at least 1,867 ha. Previous studies have recommended similar minimum population sizes but have suggested that much larger areas of suitable habitat are required (3,000–4,000 ha; Toepfer et al. 1990; Temple 1992). Temple (1992) recommended that a population size of 280 grouse would be needed in Wisconsin but that multiple populations were required to avoid issues relating to genetic, demographic, and environmental stochasticity. Similar to our results, Temple (1992) found that the extinction probability rose sharply after the population dropped below 200 individuals. Toepfer et al. (1990) recommended a minimum population size of 200 birds based on previous translocations that were poorly documented.
Although our estimates of a minimum viable population were similar to those in previous studies, our estimates of the minimum area required were generally lower. Temple (1992) recommended that at least 4,000 ha of suitable habitat was required, whereas Toepfer et al. (1990) suggested that populations be maintained in areas of at least 3,000 ha, composed of at least 1,000 ha of undisturbed grass–shrub habitat. Our estimate is more comparable to the smaller estimate of Toepfer et al. (1990), which represented the amount of undisturbed habitat required within a larger matrix. The MDA from our analysis was based on estimates of population density in ideal habitat and thus represents the minimum area required if habitat is consistently good across the reintroduction sites. If there is a significant gradient in habitat quality, with smaller areas of high-quality habitat embedded in a larger matrix of lower-quality habitat, the minimum area required to support a sharp-tailed grouse population will be much higher. The MDAs are directly related to the average densities expected for a particular landscape. Reported sharp-tailed grouse densities vary from 0.4 to 25 birds per km2, with a median value of approximately 5 birds per km2 (Table S1, Supplemental Material). Recalculating the minimum area required using the range of reported densities suggests that the MDA varies from 1,120 to 70,000 ha. Using lower population densities significantly increases estimates of the minimum area required and places them more in line with those from previous studies. Although increasing the quantity of habitat can increase the overall K of birds, grassland and shrubland habitats in western Montana are generally confined by topography and human development; therefore, improving the quality of habitats already available to increase bird densities may be a more realistic management goal.
Our results suggest that a population of 280 individuals may be able to withstand the expected demographic and environmental variability to produce a viable population over a 50-y time period; however, genetic factors should also be considered over a longer time frame. Although demographic and environmental stochasticity can be more critical in the short-term, genetic variation may be the decisive factor determining whether a population persists (e.g., >50 y) and is capable of adapting to a changing environment (Lande and Shannon 1996). The projected MVP of 280 individuals lost approximately 20% of its genetic diversity in 50 y. If genetic diversity is of concern, managers should consider supplementation (such as scenarios B and C) or maintaining more quality habitat that can support a larger population (i.e., 500 individuals) of sharp-tailed grouse.
Based on reported associations between habitat improvements and prairie-grouse demography reported in the literature, our simulations suggested that comprehensive habitat management, including improvements to both nesting habitat conditions and winter shrub cover that increase nest and female survival, respectively (scenarios I and J), would be required to produce a viable population of sharp-tailed grouse. Although our results do not represent absolute rates of extinction, they do indicate that relative to other management actions and translocation techniques, a combination of nesting and winter habitat improvements will provide the best opportunity for producing a sustainable sharp-tailed grouse population. In a similar study of a reintroduction of hihi Notiomystis cincta, an endangered bird species in New Zealand, Armstrong et al. (2007) successfully identified effective management actions by using population modeling and found that a combination of several intensive habitat improvements increased the population growth rate more than one habitat improvement project alone. Likewise, several studies of translocations have found that having quality habitat was the most important factor determining the success of the project (Griffith et al. 1989; Toepfer et al. 1990; Terhune et al. 2006). For example, in a translocation of northern bobwhite Colinus virginianus, the quantity and quality of habitat were the most critical determinants of the project's success and relocating bobwhites to poor or submarginal habitats was not recommended (Terhune et al. 2006). For sharp-tailed grouse, pre- and post-translocation management efforts should focus on improving or maintaining quality nesting and winter habitat through grazing management and shrub cover improvements.
With regard to translocation strategies, supplementation alone is unlikely to maintain a population of sharp-tailed grouse in western Montana. However, our results indicate that supplementation used in conjunction with habitat improvements could increase population persistence and genetic diversity more than just habitat improvements alone. Also, translocation of only yearlings did not greatly improve population persistence but could result in a viable population if implemented along with habitat improvements. Similar to our findings, periodic supplementation reduced the loss of genetic diversity in reintroductions of capercaillie Tetrao urogallus in southern Scotland (Marshall and Edwards-Jones 1998). Supplementation of a pair of capercaillie every 5 y for 50 y after reintroduction resulted in a predicted viable population. Similarly, we observed improved genetic diversity when the restored population was supplemented periodically with 10 females.
Although our results have important implications for reintroductions of sharp-tailed grouse, there are several caveats of PVAs that are relevant to management. First, PVAs are only as good as the data that are used to build the underlying demographic models (Beissinger and Westphal 1998). Although we based our demographic rates in this study on the best available information from published literature, they were not collected from sharp-tailed grouse populations at potential reintroduction sites. Therefore, they do not represent actual habitat conditions at potential restoration sites, and the population dynamics realized in a reintroduced population may be quite different. Furthermore, using average demographic rates from the published literature should result in a viable population when study populations are selected without regard to population status, but even our baseline scenario declined and had a very low probability of persistence over 50 y. This suggests that reported demographic rates may have come largely from declining populations, which are often the primary focus of intensive studies. Therefore, the rates used here may not represent the full range of sharp-tailed grouse population dynamics. In addition, not all the demographic rates necessary for this study were available for sharp-tailed grouse. The demographic rates used as a substitute included those from similar species, including greater sage-grouse and prairie-chickens, which may exhibit different population responses to the translocation and habitat management scenarios examined here. Thus, our estimates of population viability are based on the best available information, but true population viability may be different. Nevertheless, the relative performance of our considered scenarios should be unaffected by potential downward biases in reported vital rates because they were consistently applied. Reassessing viability as more information is available about sharp-tailed grouse in Montana will be important for better estimating the true sustainability of a reintroduced population.
Stochastic single-population models, such as the model presented here, require estimates of variance that are typically difficult to obtain (Beissinger and Westphal 1998). However, it is important to include estimates of variability or estimates of population persistence will be biased upward. The effects of environmental stochasticity are evident in the model outputs of all scenarios, and both demographic and environmental variations cause sharp rises and falls in population size over time (Figure S2, Supplemental Material). The effects of stochasticity are particularly relevant to small populations where there are fewer individuals to act as a cushion when population growth rates fall (Shaffer 1981). Incorporating estimates of environmental variation into subsequent sharp-tailed grouse population modeling and management decisions will be important for providing useful information on population viability.
Finally, PVAs as implemented in VORTEX have several limitations regarding the complexity of models that may restrict their ability to realistically represent natural systems (Reed et al. 2002). These limitations include not incorporating individual variation, not being spatially explicit, and assuming the environment is relatively static. Failure to include individual heterogeneity assumes that fates of all individuals are generally similar and can significantly overestimate the importance of demographic stochasticity and, consequently, extinction risk (Fox and Kendall 2002). Nevertheless, mean generation time for female sharp-tailed grouse is 1.5 y (Sisson 1976), so the inclusion of individual variation would likely have relatively little effect on our estimated probabilities of extinction. The assumption regarding a static environment over time and space may have more serious implications for the applicability of PVA results. A spatially implicit model does not allow for gradients of demographic rates across habitat quality. Our estimates of the MDA are based on population densities in ideal habitat and assume that the habitat quality is consistent within restoration sites. Thus, if there is significant variation in habitat quality within a restoration area, population projections will likely be overly optimistic. In addition, the model assumes that habitat conditions will not change over the time period for which population persistence is projected, which does not account for future habitat degradation or improvement.
Habitat evaluation and improvement
Based on the results of our PVA, we recommend that sharp-tailed grouse nesting and winter habitat be the focus of pre- and postrelease management. Before reintroduction, habitat surveys are suggested to adequately quantify the amount of suitable habitat available and identify target areas for improvement. If quality nesting habitat and winter habitat are lacking, management aimed to improve grazing practices and improve winter shrub cover in target areas before reintroduction can increase the probability of population persistence. Also, an assessment of both habitat conditions and the impacts of potential management actions from sharp-tailed grouse populations that could be potential source populations may provide insight into management at reintroduction sites if conditions are similar.
Postrelease monitoring of sharp-tailed grouse habitat and demographic rates to guide management decisions is vital to population recovery and future management success. Postrelease monitoring will allow for estimates of population parameters based on the dynamics of sharp-tailed grouse at actual reintroduction sites and their relationships with management actions, which will allow for the identification of the most effective management actions for specific sites. We based our management scenarios on population changes from species other than sharp-tailed grouse, which inevitably introduces a degree of uncertainty. In situations where there is uncertainty regarding the consequences of management actions, an adaptive approach, based on postrelease monitoring of parameters such as abundance or fecundity, could be a useful tool. Adaptive management involves both the development of predictive models and the subsequent updating of those models and related management and has been effective at facilitating species recovery in the past (Armstrong et al. 2007). Management based on population monitoring, evaluation, and manipulation will be important to the long-term success of sharp-tailed grouse populations in western Montana.
Please note: The Journal of Fish and Wildlife Management is not responsible for the content or functionality of any supplemental material. Queries should be directed to the corresponding author for the article.
Figures in Supplemental Material display stochastic population model outputs for each scenario of sharp-tailed grouse Tympanuchus phasianellus management in western Montana, projected over 50 y by using the program VORTEX 10. We estimated population model parameters using the best available information from published literature on prairie-grouse demographic rates during 1939–2016. Scenario descriptions and further results can be found in Tables 3 and 4.
Figure S1. Mean population size (N), mean genetic diversity, and probability of population persistence of sharp-tailed grouse Tympanuchus phasianellus populations under reintroduction scenarios with a carrying capacity (K) of 500 in western Montana, projected over 50 y.
Scenarios represent translocating yearling grouse only (A), translocating 10 adult (Ba) and yearling (Bb) female grouse every year for 5 y after reintroduction and the genetic rescue of 10 adult (Ca) and yearling (Cb) female grouse every 10 y, removal of ravens Corvus corax as nest predators (D), decreased conifer cover (E), improved grazing practices (F), increased shrub cover (G), decreased conifer cover and increased shrub cover (H), and improved grazing practices and increase shrub cover (I). The baseline scenario is shown for reference.
Found at DOI: https://doi.org/10.3996/112017-JFWM-090.S1 (704 KB JPG).
Figure S2. Population size (N) of sharp-tailed grouse Tympanuchus phasianellus populations under reintroduction scenarios with varying carrying capacities (K) in western Montana, projected by 1,000 iterations over 50 y. Each line represents one iteration. Population sizes are not specified but converge on the K if the population does not fall to zero within the 50-y period. Scenarios represent translocating yearling grouse only (A), removal of ravens as nest predators (D), decreased conifer cover (E), improved grazing practices (F), increased shrub cover (G), decreased conifer cover and increased shrub cover (H), and improved grazing practices and increase shrub cover (I). The baseline scenario is shown for reference.
Found at DOI: https://doi.org/10.3996/112017-JFWM-090.S2 (3.18 MB JPG).
Figure S3. Population size (N) of sharp-tailed grouse Tympanuchus phasianellus populations under reintroduction scenarios with varying carrying capacities (K values) in western Montana, projected by 1,000 iterations over 50 y. Each line represents one iteration. Population sizes are not specified but converge on the K if the population does not fall to zero within the 50-y period. Scenarios represent translocating 10 adult (Ba) and yearling (Bb) female grouse every year for 5 y after reintroduction, genetic rescue of 10 adult (Ca) and yearling (Cb) female grouse every 10 y, improved grazing practices and increased shrub cover (I), and a combination of supplementation of 10 adult female grouse every 10 y with both improved grazing practices and increased shrub cover (J). The baseline scenario is shown for reference.
Found at DOI: https://doi.org/10.3996/112017-JFWM-090.S3 (1.85 MB JPG).
Table S1. Densities of sharp-tailed grouse Tympanuchus phasianellus populations in the breeding and nonbreeding season reported during 1939–1976 in the United States.
Found at DOI: https://doi.org/10.3996/112017-JFWM-090.S4 (546 KB DOCX).
We thank C. Hammond and A. Wood (Montana Department of Fish, Wildlife and Parks) for initiating this work and for guidance and B. Cascaddan, A. Hicks-Lynch, A. Netter, S. Otto, J. Payne, S. Vold, and S. Wyffels for discussions that informed this study and for reviewing the manuscript. We thank J. Cummings, D. Musil, one anonymous reviewer, and the Associate Editor for comments that improved the manuscript. Montana State University and the Montana Agriculture Experiment Station provided funding and facilities.
Any use of trade, product, website, or firm names in this publication is for descriptive purposes only and does not imply endorsement by the U.S. Government.
Citation: Milligan MC, Wells SL, McNew LB. 2018. A population viability analysis for sharp-tailed grouse to inform reintroductions. Journal of Fish and Wildlife Management 9(2):554–570; e1944-687X. doi: 10.3996/112017-JFWM-090